Reactive oxygen species (ROS) may be either compounds with free radical character containing at least one unpaired electron, e.g. triplet (ground) state oxygen e02), hydroxyl (.OH), superoxide anion (02'-), peroxy and alkoxy (R02' and Ra") radicals, or non-radicals such as singlet oxygen eOz) and hydrogen peroxide (H202). They play essential roles in biochemical pathways, food degradation and in disease. Their generation is associated with processes such as senescence and pathogen attack, but they may also be a consequence of extreme environmental conditions, e.g. high or low temperature, herbicides, air pollutants, UV irradiation, nutrient deficiencies, toxic metals. (Smimoff, 1993) Three biochemical systems, which are used in this work to oxidise phenolic compounds, are described in more detail.
02'- is generated in nearly all aerobic cells and can cross membranes through a specific 'channel', whereas .OH which is formed in different parts of the cell, cannot diffuse away from its site of formation since it reacts with virtually every component it meets.
Xanthine oxidase which occurs in milk, liver and jejunum, catalyses the transformation of xanthine to uric acid generating O2'- (Equ. 2.1.; Terada et aI., 1990).
xanthine
uric acid Equ.2.1.
A main route for .OH generation is via the Fenton reaction (Equ. 2.2.) where H202 oxidises ferrous (Fe2+) ion to ferric (Fe3+) ion. Since the presence of H202 and a small amount of Fe2+is normal in vivo, the Fenton reaction is cornmon. The .OH radical reacts in three ways: by abstraction of a hydrogen atom to form H20, addition to another structure, e.g. aromatic rings, and acceptance of an electron, e.g. from the chlorine ion.
Equ.2.2.
The system horseradish peroxidase (HRP) and H202 is an alternative for oxidising phenolic compounds. Fig. 2.13. shows a generalised reaction scheme for heme peroxidases and catalases (Jakopitsch et aI., 2005). HRP, a ferric protoporphyrin IX, which is bound ionically or covalently to cell-wall polymers and is localized in the apoplastic space, reacts with H202 to form three compounds. Compound I (a ferryl porphyrin 1t-cation radical or a ferryl protein radical) is generated by oxidation of the native enzyme with one H202-molecule. It can either react directly back to the ferric enzyme by another H202 or indirectly via compound II (a ferryl species or protein radical) by two one-electron reductions in the presence of an one-electron donor (e.g.
phenolic compounds). Addition of O2'- then generates the so-called compound ill(a ferrous-dioxy/ferric-superoxide complex). A physiologically relevant way for its formation is reaction of the enzyme with 02'- produced by the oxidative cycle with a suitable substrate such as NADH (Chen and Schopfer, 1999). Compound ill is also formed from compound II with an excess of H202 or from the ferrous heme protein by dioxygen binding.
The formation of .OH radicals is also related to superoxide anion production in the cell.
Superoxide anion radicals dismutate to H202 and 02 in the presence of superoxide dismutase (SOD) (Equ. 2.3.). The so-called Haber-Weiss reaction (Equ. 2.4.) of superoxide and H202 leads then to .OH radical generation. (Smirnoff, 1993)
Equ.2.3.
Equ.2.4.
Free radicals can be detected by VarIOUS biochemical methods, e.g. detection of superoxide anion radicals by reduction of cytochrome c (Green and Hill, 1984), detection of OH-radicals by a tluorimetric detection of the hydroxylation of benzoate or by a photometric detection of the degradation of deoxyribose (Chen and Schopfer,
1999). All these biochemical methods are indirect determinations of the radicals, where the sample matrix has to be treated and hence changed in a special way before
detection. Since radicals are often highly reactive compounds, the actual status in the samples may not be reproducible any more. Such methods have to be taken with care and under consideration of possible errors in the results.
O~
O"
2
+'R-PorFe'v=O Compound I
Compound III
I
R-PorFeIlI-02"I
*
Fig. 2.13. Reaction scheme of heme peroxidases and catalases. (Jakopitsch et al., 2(05).
(AH2 - one-electron donor)
2.4.1. EPR Spectroscopy
Electron paramagnetic resonance (EPR) or electron spin resonance (ESR) spectroscopy is a non-invasive method for the detection of paramagnetic molecules, such as free radicals and many transition-metal ions (Bolton. 1972). The physical condition of the
from the 2003 lectures of Prof. Vana (Vienna University of Technology), and from Goodman & Raynor, 1970, Bolton, 1972, and Poole, 1997.
Princivles of EPR
The principles of EPR can be described by quantum-mechanical theory. An electron is a charged particle in motion and creates a magnetic moment which is described by the spin quantum number ms• (this has the values of :t112). Inthe absence of a magnetic field the energies of the spin states are equal. Applying an external field leads to separation of energies and a change in population of the energy states according to the Boltzmann statistics. This energy separation is called the electronic Zeeman effect (Fig.
2.14.). The energetically more favourable state has the electron magnetic moment parallel to the external field. The energy states are equal to gJ.lBBoms where g is the g-value or g-factor, a specific constant typical of the molecule containing the unpaired electron, J.lB is the Bohr magneton (9.27.10.24 1rr), Bo the external magnetic field, and ms the electron spin quantum number (:tl/2). The energy separation (LlE) grows with an increasing magnetic field. Transitions between the electron spin energy levels can be induced by the absorption of electromagnetic radication when the energy of the photons hvis equal to LJE (Equ. 2.5., where h is the Planck's constant (6.62607.10,24 1s) and v is the microwave frequency).
Equ.2.5.
Spectrometers operate in various microwave frequency bands, L (-1GHz)-, S (3 GHz)-, X (9 GHz)-, K (24 GHz)-,
Q
(34 GHz)- and W (90 GHz)-band, though X-band is the most commonly used one.Energy
ms
=
+ 1/2ms
= -
1/2Bo magnetic field (B)
Fig. 2.14. The electronic Zeeman effect including the resonance conditions. (Poole, 1997)
EPR spectra
EPR spectra are obtained by measuring the absorption of the microwave energy at the resonance conditions in Fig. 2.14. Normally the magnetic field is scanned while the frequency is fixed and spectra are obtained as the first derivative of the absorption curve (Fig. 2.15.). Resolution of overlapping peaks can be improved by recording the second derivatives. An EPR spectrum is characterised by the g-value, the line width and shape, the signal intensity, the hyperfine splitting and the anisotropy.
il-value
The g-value is the proportionality constant of Equ. 2.5. and is equal to 71.44775.vIB0, where v is in GHz and Bo is in T. For a free electron g = 2.0023. The g-values of different molecules differ from that of the free electron because of the interaction of the electron spin and orbital angular momenta, which leads to a shift in the resonance energy «2.0023 if coupling is with empty orbitales), >2.0023 if coupling is with filled orbitals). g-values of solid samples are often anisotropie, which means they are
a
b
c
Fig. 2. I 5. EPR-spectra shown as absorption peak (a), the first derivative (b) and the second derivative of the absorption peak (c).
Relaxation, line width and shave
The resonance situation leads to two opposing processes in the molecule. On one hand the microwave field tries to equalise the difference in the population density, whereas relaxation processes try to restore the Boltzmann distribution. The latter consist of 2 separate processes - spin-lattice-relaxation and spin-spin-relaxation.
Ina conventional EPR experiment, spin-lattice-relaxation is responsible for maintaining a constant absorption signal since the interaction between the electron spin and the lattice re-establishes the Boltzmann distribution. Without this interaction the electron spin energy levels would rapidly become equally populated; then no further microwave energy could be absorbed, the lines become broader and the transition becomes saturated. On the other hand a broader absorption line also occurs when the relaxation
time is too short, which can be explained by the Heisenberg uncertainty principle - a short relaxation time (Tt) leads to a broader energy state (~v
=
lffl). The interaction between the spin and the lattice occurs through two processes - the direct and the Raman process. The direct process involves the exchange of a complete quantum Vowith a lattice vibration, the Raman process takes place over a two-photon transition, where an additional photon Vt is absorbed first to the available energy state and another one V2is then emitted. The difference between Vt and V2is the initial frequency Vo.This process occurs only when enough photons are available which is the case at high temperatures, whereas the direct process dominates at very low temperatures. The relaxation time is determined as the time from saturation to the recovery of the thermal equilibrium.
The spin-spin-interaction occurs between the spins of unpaired electrons. The movement of the spin in the external magnetic field can be considered as a magnetic dipole having a fixed component in the direction of the magnetic field; this produces an additional field at a neighbouring unpaired electron, resulting in a shift in the total field and hence a shift in its energy levels. This interaction is usually seen in the spectrum as line broadening. Additionally the broadening of the absorption takes place for electrons in neighbourhood which have the same Larmor frequency (equal to the same g-value) since the oscillating field induces transitions in the adjacent electron leading to a decrease of the normallife time.
A normal dipole-dipole interaction leads to a Gaussian line shape. Interactions between spin and lattice or motional averaging effects will narrow the lines - exchange narrowing - resulting in a Lorentzian shape. The line width represents the energy distribution within the energy levels and the interaction of the unpaired electron with its environment. Ifhyperfine splitting constants or the separation of different components in a sample are smaller than the line width, the absorption line will broaden.
Sümal intensitv
•
•
the radicals and their high reaction rate. However, there are semiquantitative methods using standards like DPPH or Mn which are run at the same time as the samples.
Hvverfine svlittinf!
In addition to the external magnetic field, nuclei with non-zero spins such as IH (1=1/2),
14N (1=1), 13C(1=1/2) create additional magnetic fields which interact with the unpaired electron and cause a splitting of the energy levels into 21
+
I components. This leads to a splitting of the peaks in the EPR spectrum which is called the hyperfine splitting. The change of the nuclear spin quantum number ml is much slower than the one of the electron spin quantum number ms, therefore ml is fixed during an electronic transition and two possible transitions are allowed according to the selection rules of ms=
:tl andml
=
0 (Fig. 2.16.). A typical hydrogen atom spectrum where two peaks of equal intensityare visible is given below the energy diagram. The distance between the peaks is called the hyperfine splitting described by the hyperfine coupling constant, a (which is usually quoted in mT or gauss). There is a direct proportionality between the hyperfine splitting and the product of the magnetic moment f..I.N of the nucleus and the fractional occupancy of the molecular orbital containing the unpaired electron.The interaction of the unpaired electron can be with more than one nucleus of the same type, e.g. with hydrogen atoms from a CH-, CH2- or CH3-grouP resulting in an intensity relation according to Fig. 2.17. Interactions with different type of nuclei lead to spectra showing the strong interaction with one type of nucleus leading to (211
+
I )-splittings and all these lines are further split due to the weaker interactions with the other nucleus into (2h+ I)-splittings. No overlapping sets of hyperfine splitting lines occur if the weaker interaction is very small compared to the strong one.I I I I
-~
Ia
n fi
I\
ii
: iV
H=O
•
Fig. 2.16. Separation of the energy levels dependent on the magnetic field shown in the hydrogen atom. The dotted transition indicates the situation where a would be zero.
(Bolton, 1972)
•
1=112 n=11 1
2
1 2 1
3
1 3 3 1
4
1 4 6 4 1
5
1 5 10 10 5 1
•
Anisotropy
The hyperfine splitting due to the interaction of the electron with the nucleus is dependent on the type of orbital in which it occurs, e.g. s,p or d.Ifthe electron is in an s orbital, the hyperfine coupling constant will be large due to the high electron density at the nucleus and independent of direction since s orbitals are symmetrical (isotropic hyperfine splitting). Anisotropic hyperfine splitting occurs when the electron is in ap or d orbital. Inthis case there is no electron density at the nucleus, the interaction which is based on two magnetic dipoles, is small and dependent on the direction (resolved inx, y, and z) of the orbital relative to the applied magnetic field and to the separation of the dipoles. The magnitude of this hyperfine coupling is zero when it is integrated over all directions. However, a small isotropic hyperfine splitting is usually observed as a result of polarisation of filled s-orbitals by the unpaired electron(s). (see e.g. Goodman and Raynor, 1970) Hybridisation of s, p, and d orbitals leads to a combination of isotropic and anisotropic couplings.
There is also a dependency of the hyperfine splitting on the physical state of the sample.
Any solid matrix including frozen solutions will show the sum of isotropic and anisotropic interactions whereas with fluid solutions only the isotropic coupling will be observed due to the fact that anisotropic coupling is averaged to zero. In the case of large biochemical molecules in fluid systems, the tumbling frequency of the molecule may be lower than the resonance frequency leading to an anisotropic spectrum similar to that associated with a solid state molecule. This situation is also often observed when spin traps are used and large molecules are trapped leading to line widthlheight variations in the spectrum.
Saturation
The Boltzmann equation (Equ. 2.6.) describes the number of electrons in the ground (Ngd) and excited states (Nex), where L1E is 2gp.JJ. Resonance can only be observed if Nex is different to Ngd. Any change in the occupation of the two levels is given by Equ.
2.7.
Ngd
=
e ~N~x
k Boltzmann constant (1.380622.10.23 J/K) T temperature of the system
Equ.2.6.
Equ.2.7.
Wgd-ex probability for the electron transition from the ground to the excited state Wex-gd probability for the electron transition from the excited to the ground state
Saturation occurs when Nex approaches Ngd• When Nex
=
Ngd an electron is emitted for everyone absorbed. Then an increase in the microwave power has no influence on the signal intensity. At low temperature saturation occurs relatively easy. A typical saturation curve is given in Fig. 2.18.16 .A,-'AA'-'--A'--_A .
k~ Lr ~_.-._.-._.~
.£14 p
~12
l
c '
.- 10 ' ] 8
Il
u
~ 6
A
~ 4
~ 2
o o
0.2 0.4 0.6 0.8 1.0 1.2 1.4 Square root microwave power (mW~) Fig. 2.18. Typical saturation curve of DPPH at 77 K.EP R -soectrometer
Source
I ! Magnet system tuner
I 11-"-"-"-"-"-"-"-"-'
I Automatic i i Magnet
I frequency
II
powerFig. 2.19. Diagram of a typical X-band EPR spectrometer. (Bolton, 1972)
The magnetic field is produced by an electromagnet. The production of microwaves in the EPR spectrometer we used (Broker ESP 300E, Bruker Biospin, Rheinstetten, Germany) is made by klystrons but many instruments nowadays use Gunn diodes as microwave sources. Transmission of the radiation from the source to the resonator as well as to the detector occurs with wavelength conductors. The resonator (cavity) is the part of the microwave conductor where interaction with the sample occurs and where a standing wavelength is produced. Silicium-Diodes are used for detecting the microwave signal.
Princivle of measurement
The microwave bridge is the core of the spectrometer and the connection between the microwave source and the detector. To receive high signaJ-to-noise ratios and baseline
stability, the magnetic field is modulated at low frequency. Therefore the microwave signal is also modulated with the same frequency and the resulting spectrum is obtained as a ISI derivative of the absorption. In the present work, the best value of the modulation frequency to compromise noise and resolution was tOO kHz.
Variable oarameters
Some parameters have to be chosen and set independent of the sample before the EPR spectrum can be recorded. The optimum microwave power varies a lot with the temperature at which the spectrum is recorded, e.g. frozen samples get saturated very easy; therefore the power needed is low (about 0.1 mW), freeze-dried biological samples can be recorded at room temperature with 1 mW, whereas good spectra can be obtained with spin-trapped samples using 10 or 20 mW microwave power (Pirker, 2002). The choice of the modulation amplitude is a compromise between the signal height and peak resolution. With decreasing modulation amplitude, resolution is improved, but the peak height is decreased. Another parameter is the receiver gain. It should be chosen to be big enough to obtain a good signal considering that an increasing receiver gain also increases the noise level. The speed of recording one spectrum can be regulated by the conversion time whereas the time constants filter out the noise by slowing down the response time of the spectrometer. The centre field and the sweep width of a spectrum are two other parameters which are adjusted to the sample conditions.
All simulations of EPR spectra were carried out using the program SimFonia from Bruker.
2.4.2. Spin Traps
The generation of unstable short-lived free radicals in a system can be detected by using molecules called spin traps. The technique was first demonstrated in the late 1960s (Janzen, 1984). Spin traps are diamagnetic compounds that react with radicals to produce new long-lived radicals which are stable enough to be detected by EPR.
Qualitative information can be obtained from the parameters of the spectrum. An unknown trapped compound can be classified as belonging to a specific group of radicals, e.g. carbon-centred, oxygen-centred, or sulphur-centred, as a result of its EPR parameters (The older literature was collected in the N.I.E.H.S. spin trap database, but this has not been maintained in recent years). Since solvent effects can have major effects on the hyperfine splitting of the spin adduct, the solvent used in the experiment has to be clearly stated. (Buettner, 1987) The most commonly used spin traps are nitrones and nitroso compounds. In this work some derivates of dimethyl-I-pyrroline N-oxide (DMPO) such as 5-(diethoxyphosphoryl)-5-methyl-I-pyrroline N-oxide (DEPMPO), 5-(di-n-butoxyphosphoryl)-5-methyl-I-pyrroline N-oxide (DBPMPO), 5-(bis-(2-ethylhexyloxy)phosphoryl)-5-methyl-I-pyrroline N-oxide (DEHPMPO), and 5-(di-n-propoxyphosphoryl)-5-methyl-1-pyrroline N-oxide (DPPMPO) were used as well as the spin traps phenyl-N-t-butylnitrone (PBN), a-(4-pyridyl-I-oxide)-N-t-butylnitrone (4-POBN), 5-propyloxycarbonyl-5-ethylpyrroline-1-oxide (PEPO), and 1,3,3-trimethyl-6-azabicyclo[3.2.1 ]oct-6-ene-N-oxide (TRAZON). Their structures are given in Fig.
2.20. The radical attack takes place at the unsaturated a-carbon atom next to the nitrogen atom producing an adduct called nitroxide radical. The unpaired electron is delocalised over the molecular orbital which is the reason for the stability of the radical.
All spin traps except PBN and 4-POBN were synthesised at the Research Institute of Biochemical Pharmacology and Molecular Toxicology at the University of Veterinary Medicine, Vienna.
The DM PO derivates have been used extensively with biological systems (see e.g.
Buettner, 1987). However, DMPO is unable to discriminate between the superoxide radical anions and the hydroxyl radicals. This problem has been overcome with various phosphorylated derivates, which are suitable for studying the generation of superoxide radical anions and peroxyl as well as hydroxyl radicals (Fréjaville et al., 1995). The traps used were designed to probe selectively either the aqueous or lipid phase of the samples.
DEPMPO is very hydrophilic. Increasing the size of the organic chain on the phosphoryl group increases the lipophilicity of the spin trap. The EPR spectra of the phosphorylated derivatives of DMPO consist of two doublets from interaction of the unpaired electron with nuclear spins of the isotopes 31p and IH (both having I
=
Y2)and a triplet from the interaction with the spin 14NCI
= 1), giving a total of 12 peaks if the hyperfine splitting constants of 31p, IH, and 14N are not equal (Fig. 2.21.). The synthesis of DMPO derivates is given in Scheme 1 (Stolze et al., 20(0).PEPO is a highly hydrophilic spin trap, which has been reported to form exceptionally stable superoxide adducts (half-life c. 20 minutes, Stolze et al., 2(03). Spectra are simpler than with the phosphorylated spin traps described in the previous paragraph, and consist of at least 6 peaks with just one IH and 14N nucleus contributing to the resonance (Fig. 2.22.). The synthesis (Stolze et aI., 2(03) is described in Scheme 2.
The final spin trap synthesised at the University of Veterinary Medicine, Vienna, was the bicyclic nitrone, TRAZON, which forms relatively stable adducts with lipid alkoxyl radicals (Stolze et aI., 2002). The spectra are often quite complex, since both exo and endo adducts can be formed and each of these has hyperfine structure from up to five IH
The final spin trap synthesised at the University of Veterinary Medicine, Vienna, was the bicyclic nitrone, TRAZON, which forms relatively stable adducts with lipid alkoxyl radicals (Stolze et aI., 2002). The spectra are often quite complex, since both exo and endo adducts can be formed and each of these has hyperfine structure from up to five IH